Synthetic survey catalogs for the Galactic Roman Infrared Plane Survey (GRIPS) using py-ananke

The Nancy Grace Roman Space Telescope, named in honor of NASA’s first Chief of Astronomy, is a next-generation observatory designed to answer critical questions in cosmology, exoplanet exploration, and infrared astrophysics. Scheduled for launch in the mid-2020s, the Roman Space Telescope is poised to revolutionize our understanding of the universe.

Equipped with a 2.4-meter (7.9 feet) primary mirror, the Roman Space Telescope boasts a field of view 100 times greater than that of the Hubble Space Telescope, while maintaining similar resolution. Its Wide Field Instrument (WFI) will capture vast regions of the sky in unprecedented detail.

One of the key features of the Roman Space Telescope is its capability in near-infrared photometry. This ability allows it to peer through dust and gas clouds that obscure our view in visible light, making it an ideal tool for exploring the heavily extincted galactic plane. For that reason, the Roman early definition science team selected a galactic plane survey as top priority.

As such the Roman Space Telescope will play a crucial role in advancing galactic science. By surveying large areas of the galactic plane, the telescope will create detailed maps of the distribution of stars, star-forming regions, and the structure of the Milky Way.

The Roman Infrared Nearby Galaxy Survey (RINGS), led by Principal Investigator B. Williams, is a major Roman Large Wide Field Science Program. It aims to develop advanced software tools for simulating and analyzing Roman surveys of nearby galaxies. This project was approved in response to the NASA ROSES solicitation "D.14 Nancy Grace Roman Space Telescope Research and Support Participation Opportunities" (number NNH22ZDA001N-ROMAN).

RINGS aims to provide the scientific community with robust software tools that generate simulated Roman imaging data based on numerical simulations. By simulating Roman observations, the community can validate and refine model predictions related to star formation, galaxy structure, and dark matter distribution. Moreover, these tools will help optimize observing strategies, allowing us to efficiently plan Roman observations of resolved stellar populations in both crowded and low surface brightness regions.

The challenges of source separation in crowded fields are shared by both the RINGS project and galactic plane surveys. By addressing these challenges, the RINGS project will not only advance our understanding of nearby galaxies but also enhance the analysis of data from galactic plane surveys. This synergy will improve our ability to interpret complex stellar environments, benefiting a wide range of astronomical research.

Py-ananke (Thob et al. 2023) is an advanced synthetic survey tool developed as part of the RINGS crowded-field photometry pipeline. The software offers a comprehensive pipeline designed to generate mock astrometric and photometric catalogs of synthetic stars derived from particle-based simulated star populations. Developed as a Python package, py-ananke is compatible with Python 3.8, 3.9, and 3.10.

Currently in beta, the software has been submitted to the Journal of Open Source Software (JOSS) and is currently under review. Despite this, you can readily access its pre-release in its main GitHub repository from where it can be installed and tested. We encourage users to utilize the Github issues feature to report any bugs encountered or suggest new ideas for improvement. Contributions to the project are highly valued and greatly appreciated.

The development of py-ananke was driven by the need to make the sophisticated framework of ananke more accessible and versatile. Originally described in Sanderson et al. 2020, ananke focused on turning cosmological simulations, such as the Latte suite of FIRE simulations, into synthetic GAIA catalogs. But these simulations have limited resolution and cannot accurately represent fully resolved stellar populations with individual stars.

To address this limitation, the authors of ananke developed a framework of scripts and data that enabled the generation of synthetic GAIA star surveys from these simulated galaxies. This framework combines density estimations and Initial Mass Function (IMF) sampling techniques to create representative populations of mock stars.

py-ananke enhances this framework in three key ways:

• Ease of Installation and Distribution: py-ananke is a self-contained and easily installable Python package, facilitating the usage and adoption of the ananke framework by a broader community.
• Expanded Photometric Systems Library: The package supplements its library of photometric systems, augmenting its versatility to simulate a larger number of facilities, including the Roman Space Telescope.
• Modular Implementation: By modularizing its implementation, py-ananke allows for easier maintainability and upgradability of the package, ensuring that it can evolve with the needs of the scientific community.

These enhancements make py-ananke a powerful tool for generating and analyzing synthetic survey data, advancing our ability to simulate observations from the Roman Space Telescope and other facilities.

The implementation of py-ananke is designed to streamline the ananke pipeline, and to prevent the need for the user to manually handle the interface between Python and the C++ backend software.

It notably introduces dedicated wrapper submodules (hosted in repositories that are separate from that of py-ananke, but linked as git submodules), namely py-EnBiD-ananke and py-Galaxia-ananke, specifically developed to handle the installation and utilization of these C++ subroutines, namely EnBiD (Sharma & Steinmetz, 2006, 2011) and a modified version of Galaxia (Sharma et al., 2011a, 2011b) called galaxia-ananke. These submodules relieve users from the need to directly manage the C++ software while isolating the C++ wrapping process.

This allows py-ananke to focus on processing inputs and outputs using pure Python. Figure 4 illustrates the inner framework process of the full pipeline, showcasing the various module and submodule classes and where they are used in an input to output fashion from left to right. The implementation of involves six classes, with only one - Ananke - being relevant to the end user:

• Ananke objects serve as the user interface, connecting all of py-ananke’s classes and the py-Galaxia-ananke classes to execute the full pipeline via its method run
• Universe objects store the particle data and various parameters provided to Ananke
• Observer objects store the observing configuration, including the position in space
• DensitiesDriver objects utilize the particle data from the Universe class to compute and store phase space densities, employing py-EnBiD-ananke
• ExtinctionDriver objects are utilized by Ananke objects for post-processing to estimate and append extinctions in the output catalogs of py-Galaxia-ananke, only if the user specified dust column densities per star particle
• ErrorModelDriver objects are utilized by Ananke objects for post-processing to determine and append errors on the quantities in the output catalogs of py-Galaxia-ananke

The latter two driver classes require respectively extinction coefficients and error models that are photometricsystem-dependents and can be specified by the user. Also, py-ananke is designed with dedicated source files to contain default implementations, which currently only hold default for the Gaia photometric system. Some unreleased updates have already  implemented an integration with the Spanish Virtual Observatory Filter Profile Service (Rodrigo et al., 2012, Rodrigo & Solano, 2020) to link with their bands database and generalize extinction coefficients using Wang & Chen 2019 equation 9 & 10 curves. Future updates will continue to expand this further.

As mentioned earlier, the Roman early definition science team recently recognized the importance of exploring the galactic plane with Roman, and ranked the need to prepare a galactic plane survey as a top priority. The Galactic Roman Infrared Plane Survey (GRIPS) is an independent initiative that is an acceptable example of such survey, as proposed by Paladini et al. (2023) in their 2021 white paper. GRIPS aims to leverage the unique capabilities of the Roman Space Telescope to explore the heavily extincted and crowded regions of the galactic plane, providing an unparalleled dataset of approximately 120 billion sources over a ∼ 1000 deg² area.

The proposed GRIPS seeks to address several key scientific goals:

• Mapping the Galactic Structure: By surveying the inner Galactic disk and bulge, GRIPS will create high-resolution maps that reveal the detailed distribution of stars, star clusters, and star-forming regions. This comprehensive dataset will offer insights into the structure and dynamics of the Milky Way.
• Star Formation and Evolution: The survey will enhance our understanding of star formation processes and the evolutionary history of stellar populations in diverse environments, thanks to Roman's superior sensitivity and resolution in the near-infrared.
• Stellar Populations: GRIPS will facilitate the study of various stellar populations, including young stellar objects (YSOs) and red clump stars, across a broader volume of the Galactic disk. This will significantly improve our knowledge of the initial mass function and stellar distribution.
• Dark Matter Distribution: By analyzing stellar motions and proper motions, GRIPS will contribute valuable data on the distribution of dark matter within the Milky Way. Roman's high angular resolution and mapping speed will enable comprehensive coverage and detailed studies of Galactic structure.

In order to prepare for the Galactic Roman Infrared Plane Survey (GRIPS), we generated a 2% prototype synthetic survey of a Milky-Way like simulated galaxy. We employed py-ananke to create this synthetic survey, utilizing data from the m12f simulation of the latte suite of FIRE zoom-in simulations (Wetzel et al. 2016, Garrison-Kimmel et al. 2017 & Hopkins et al. 2018). We chose to generate not only simulated photometry for Roman, but also for Gaia DR2 and LSST, including extinction estimates that were determined from the line-of-sight dust column densities that resulted of assuming a given gas-to-dust ratio.

Since py-ananke allows to lower the sampling of the survey by a given fraction, we chose to generate only 2% of a representative survey due to logistical constraints, including computational resources and data storage limitations. Despite these limitations, the 2% prototype survey provides a valuable dataset for refining our techniques and ensuring the robustness of our methods before scaling up to the full survey.

The synthetic survey consists of the full population that results of using all star particles within 60 kpc of the simulated galaxy center of potential, and limiting apparent magnitudes in the Roman F₁₅₈ filter to under 25.3, one of the 3 apparent photometry limits of the GRIPS specifications. The 2 other photometry limits (under 24.6 in F₁₀₆ and under 24.7 in F₂₁₃) are applied in post-processing.

In Figure 6, we present the complete scope of our 2% prototype survey using two RGB flux maps in galactocentric Mollweide projection, depicting the Gaia DR2 and Roman apparent photometry. The upper panel highlights the Gaia DR2 bands, where dust lanes are prominently visible, significantly obscuring a large portion of the synthetic stars. In contrast, the lower panel showcases the Roman near-infrared bands, which reveal a substantial number of stars that were completely hidden behind the dust in the Gaia DR2 bands.

These contrasting flux maps underscore the importance of the Roman Space Telescope's near-infrared capabilities in studying the heavily extincted regions of the Galactic plane. By penetrating the dust that obscures visible light, Roman allows us to detect and analyze stellar populations that would otherwise remain hidden. This capability is crucial for constructing a comprehensive and accurate map of the Milky Way's structure.

While Figure 6 illustrated the unrestricted survey regardless of astrometric specifications, Figure 7 focuses on the GRIPS-defined survey area. Specifically, this includes all stars with galactic longitudes and latitudes within ±60° and ±3° for the disk, and within ±10° in both coordinates for the bulge. Each panel displays a color-magnitude diagram in Gaia DR2, LSST, and Roman reddened absolute photometry, with heatmaps representing stars that fall within each instrument's detection limits in apparent magnitudes:

• m_G < 20.7 for Gaia DR2
• m_u < 23.8, m_g < 24.5, m_r < 24.03, m_i < 23.41, m_z < 22.74 & m_y < 22.96 for LSST
• m₂₁₃ < 24.7, m₁₅₈ < 25.3 & m₁₀₆ < 25.5 for Roman/GRIPS

These detection limits constrain our survey to between 22 and 25 million stars in the optical, heavily extinction-prone bands of Gaia and LSST. In contrast, the Roman Space Telescope's near-infrared capabilities allow access to an astounding 670 million stars in the survey area. Additionally, the Gaia and LSST color-magnitude diagrams include contours that highlight the extent of the stellar distribution when incorporating the stars undetected with Gaia and LSST, but detected by Roman. This comparison underscores Roman's superior ability to penetrate dust and detect stars in the galactic plane, significantly enhancing the overall survey depth and coverage.

To further characterize the differences in detection-restricted populations, Figure 8 presents additional color-magnitude diagrams but with unreddened unextincted absolute photometry. The left, center, and right panels display the same color-magnitude diagrams, with the upper panels showing stars that fall within both Roman and Gaia or LSST detection limits, and the lower panels illustrating stars that are only within Roman's detection limits. Naturally, the Roman-only population displays a large number of cooler, low-mass red dwarfs, which have more prominent photometry in the near-infrared part of their spectrum. However, they also show a non-negligible population of stars nearly identical in distribution to the upper diagrams, that spread across the entirety of the main sequence, extending all the way to the asymptotic giant branch. This highlights Roman's unique ability to detect a broader range of stellar types, particularly those obscured in the optical bands.

We have made significant progress with our current setup, and our prototype survey include the following properties:

• Roman, LSST & Gaia DR2 Photometry We have generated separate photometric data for each synthetic star, with each photometry components (intrinsic magnitudes, distance modulus, extinction, ...) stored separately.
• Simple Astrometry: The catalog has simple astrometric data as projected, without any error modeling, including parallax, celestial, galactic and galactocentric coordinates, as well as their corresponding proper motion/velocity properties
• Simulation-Carried Properties: We have included simulation-derived properties such as ages and chemical abundances, which are carried from the parent particle and stored in the synthetic catalog.

Our future work will focus on expanding and optimizing our methodologies and survey coverage:

• Incorporating Gaia DR3 and Euclid Data: We plan to include photometry simulating Gaia DR3 and Euclid filters, to allow for more preliminary galactic science in cross-facility settings.
• Optimization for Full-Scale Surveys: We aim to optimize our techniques to enable the generation of a full 100% synthetic survey, overcoming current logistical constraints, which would be a requirement to further study the effects of crowding.
• Post-Processing Enhancements: We will tweak and improve our post-processing methods to refine the data quality and analysis; this will notably include the simulation of an error model, or that of an interstellar dust infrared emission map.
• Exploring Different Simulations and Observer Positions: To broaden our understanding, we will utilize various simulations  in the latte suite, as well as different observer positions, for exploring further the impact of our position in the Milky Way.

We are keen to engage with the broader scientific community and incorporate feedback into our future work. We welcome suggestions and collaborations. What Would You Be Interested In? We are open to exploring additional areas of interest based on community input. Whether it's specific regions of the Galactic plane, different types of data analysis, or new collaborative projects, we invite your ideas and contributions.

• Figure 1: STScI & CMU
• Figure 2: RINGS team
• Figure 3:
  • left: Arpit Arora,
  • right: Sanderson et al. 2020
• Figure 4: Thob et al. 2023
• Figure 5: Paladini et al. 2023
• Figure 6 to 8: made with python and matplotlib

• Garrison-Kimmel, S., et al. (2017)
  Not so lumpy after all: modelling the depletion of dark matter subhaloes by Milky Way-like galaxies.
  MNRAS, 471(2):1709–1727
• Hopkins, P. F., et al. (2018)
  FIRE-2 simulations: physics versus numerics in galaxy formation.
  MNRAS, 480(1):800–863
• Paladini, R. et al. (2023)
  Roman Early-Definition Astrophysics Survey Opportunity: Galactic Roman Infrared Plane Survey (GRIPS)
  arXiv:2307.07642
  
• Rodrigo, C. et al. (2012)
  SVO Filter Profile Service Version 1.0
  IVOA Working Draft 15 October 2012
• Rodrigo, C. and Solano, E. (2020)
  The SVO Filter Profile Service
  XIV.0 Scientific Meeting (virtual) of the Spanish Astronomical Society
• Sanderson, R. E., et al. (2020)
  Synthetic Gaia Surveys from the FIRE Cosmological Simulations of Milky Way-mass Galaxies.
  ApJS, 246(1):6
• Sharma, S. and Steinmetz, M. (2006)
  Multidimensional density estimation and phase-space structure of dark matter haloes.
  MNRAS, 373(4):1293–1307
• Sharma, S., et al. (2011a)
  Galaxia: A Code to Generate a Synthetic Survey of the Milky Way
  ascl:1101.007
  
• Sharma, S., et al. (2011b)
  Galaxia: A Code to Generate a Synthetic Survey of the Milky Way.
  ApJ, 730(1):3
• Sharma, S. and Steinmetz, M. (2011)
  EnBiD: Fast Multi-dimensional Density Estimation
  ascl:1109.012
  
• Thob, A. C. R, et al. (2023)
  Generating synthetic star catalogs from simulated data for next-gen observatories with py-ananke
  arXiv:2312.02268
• Wang, S. and Chen, X. (2019)
  The Optical to Mid-infrared Extinction Law Based on the APOGEE, Gaia DR2, Pan-STARRS1, SDSS, APASS, 2MASS, and WISE Surveys
  ApJ, 877(2):116
• Wetzel, A. R., et al. (2016)
  Reconciling Dwarf Galaxies with ΛCDM Cosmology: Simulating a Realistic Population of Satellites around a Milky Way-mass Galaxy.
  ApJl, 827(2):L23